Building an Ethernet Network for IP Surveillance

Market interest in IP video cameras is understandable. They have many advantages: a wide selection of devices, the flexibility of the software functionality, a good picture, easy integration into the computer infrastructure. It is time to think about how it is most convenient and economical to build a stable data transmission environment based on IP cameras, providing for scalability. Just throwing the optics and connecting to it the first set of media converters and simple hubs that comes to hand is also an option, but fraught with many problems in the future. It makes sense to investigate the problem a little deeper.

There are a lot of questions. What is the connection diagram? Before laying the cable, the question arises: how to put it? Which cable to lay? How many fibers should be? And what, it is also necessary to cook it? What active equipment to use? ... etc.

Let's look at all these issues in order, using the example of abstract territory.

In fig. No. 1 gives a diagram of such a territory.

Fig. 1

We omit the arrangement, choice of type and the required number of cameras, their direction, since these issues require a separate article. The perimeter of our territory is about 1550 meters. Assume that 15 IP cameras located within a radius of 100 meters from the cabinets are sufficient for video surveillance. This distance is due to the fact that the Ethernet standard regulates the operating state of a segment with a length of not more than 100 meters. Currently, the de facto standard is the use of POE technology, which allows you to supply power to the camera from the switch through the same UTP cable through which it is connected. This solves a lot of problems associated with the supply of electricity, since in this case it is enough to power the control cabinet with the switch, and you can no longer worry about powering the camera.

Thus, we get from 2 (in the diagram: a circle with a label “2k”) to 3 (in a diagram: a circle with a label “3k”) cameras per cabinet.

It would be reasonable to combine these cabinets with the server optical cable using 2 directions (shown in the diagram in red and blue). In Fig. 2, both cable directions, cabinets, and the “possible” position of cabinet No. 7, which we, for example, were not going to mount at this stage, are added to the circuit, but we want this possibility in the future.


Now the question arises: what cable design to use? The answer to this question depends largely on the method of laying. For example, if there are poles along the perimeter of the territory, it is more reasonable to use a “suspended cable with an external power element”. The design of such a cable is shown in Fig. 3.


Using the nomenclature of one of the largest suppliers of optical cable, Integra, the model of such a cable will be called IK / T-M4P-Akh. The last 2 characters mean: “A” is a single-mode cable, and the number of fibers is put in place of “x”. For example, “A8” is 8 single-mode fibers.

If the cable is planned to be laid in the ground, or mounted along the fence, it is wiser to choose a design with light armor. See fig. 4

Fig. 4

According to the same nomenclature of the Integra company, the model will be called IKSL-M4P-Ah.

In real projects, combinations of these, as well as the use of other designs are possible, but the cables described above are used most often.

The type of cable was chosen, but now the question arises, how to combine all this at the level of optical fibers, or in other words: “How will we cook, customer?”

Here it is worth considering all three possible scenarios. In short, then connect everything:

- sequentially "bus";
- a star using individual fibers / fiber per optical connection.

At this stage, the construction of the so-called "welding plan" is necessary. Why is it needed? Well, firstly, you, as a customer, will have before you a detailed wiring diagram, which in the future will be useful to you during operation. Secondly, by inviting a welding engineer from the outside, or by giving tasks to your specialist, there is no other way to clearly and clearly set the task. And thirdly, in this diagram all 3 connection options will be visible, which we will now consider.

Fig. 5

So, option No. 1: sequentially with the "bus", is given in Fig. No. 5. The black dots on this diagram indicate the places of welding, the black squares indicate the connectors, and the lines ending with a small dash are the freely left fibers.

As can be seen from the diagram, only 2 fibers will be involved in the cable, including when adding a new cabinet. If there are more than 2 fibers in the cable, it is recommended to weld free fibers at the installation stage, as shown in the example of the 3rd fiber. This will be useful in any case, since in the case of further development of the network, it will not be necessary to climb into already mounted cabinets, increasing the risk of emergencies.

Using this scheme, we see that for laying we need 2 or more fiber optical cables, and the equipment requires a minimum of 2 optical ports.

This scheme, although attractive for its simplicity, obviousness and fewer welds, however, it has one significant drawback. Imagine that something happened to the equipment in cabinet number 1. What will become of our connections in cabinets No. 2 and No. 3? Right! We will lose them.

To avoid such cases, it is necessary to continue the cable from cabinet No. 4 and return it (preferably in a different way, for example, by connecting to cabinet No. 5) to the server, thereby creating a ring. At the same time, of course, support for the “ring” by active equipment is required, and its proper configuration. And obviously, this will require an additional 420 meters of cable and managed switches, which are by no means cheap.

It should be noted that in the considered scheme, it is possible to use only one fiber, if we use WDM optical equipment, which allows transmitting and receiving a signal using 1 fiber at different wavelengths. However, this does not solve the problem described above.

An alternative to this connection scheme is to use the Star topology shown in Figure 2. 6. Fig . 6

As can be seen from this diagram, when using this topology, each cabinet will be connected "independently" from neighboring ones. Why is the word “independently” quoted? It should be understood that, of course, we will lose connections in cabinets No. 2 and No. 3, if we cut, for example, the cable between the server and cabinet No. 1. Only the construction of the real ring described above will save us from such trouble. However, from the problems with power supply or equipment failure inside the cabinet number 1, this clearly saves.

The diagram shows that the number of welds increases, because if you do not use “transit” installation, you need to weld each pair of fibers in transit through an adjacent cabinet. Of course, as in the previous version, it is possible to use WDM transceivers, which in turn will halve the number of fibers used, as well as the number of welds.

Which scheme to choose - the customer decides.

Since it is desirable for us to ensure the independent functioning of each cabinet from each other and to use inexpensive equipment, in this example we will take the star connection scheme as a basis, in which 2 separate fibers will go from the server to each cabinet.

Virtually we decided on the scheme, but how will it look in practice? Usually, in order to minimize losses on the one hand, and reliable termination on the other, it is possible to use the GP-B optical box to terminate the fibers inside the cabinets. Its appearance is shown in Fig. 7

Fig. 7

The box has two ports for the cable (input and output in our case) and a mount for the outgoing optical cord. A feature of this box is that the fibers to be terminated are welded directly to the halves of the patch cords, and the transit fibers are welded into a splice cassette. Thus, the connection is simplified (the bundle welding + pigtail + adapter + patch cord is removed), thereby reducing losses. In our case, we will use 1.5 meter halves of a 3 meter patch cord LC / UPC-LC / UPC-SMB1-DX-3M. We will address the issue of using the LC / UPC connector a little later.

In this regard, I note that some customers who want to “save” at a loss are limited only to the splice cassette and welding of the pigtail (0.9 mm in diameter), which ultimately leads to cliffs and other troubles.

This solution will reliably fix both ends of the optical cable, protect welding spots and make it possible to connect equipment using a cord protected with a 3mm sheath.

In the server issue of cable termination, the situation is somewhat different. Since the server room is often equipped with a 19-inch cabinet, in this case it is necessary to use optical cross. In our case, we need an optical cross, equipped with 16 optical ports. A good choice would be the FODF-1U-24SCSX / 24LCDX model shown in Fig. 8.

Fig . 8

This model has a lightweight aluminum case, 3 interchangeable strips designed for adapters of either SC simplex or LC duplex, and a capacious splice cassette. That’s practically all we need.

In the future, of course, we will need optical patch cords, for example, SC / UPC-LC / UPC-SMB1-DX-1M, which are perfect for connecting our equipment to this cross.

Now it is time to decide on the active equipment. Of course, the problem can be solved using office switchboards, soap dishes and media converters, thus creating a unique pile of equipment that inspires horror for maintenance engineers. Perhaps the reader has already heard, or even used the so-called "industrial" switches (like MOXA, Hirschmann, etc.). However, a solution based on them can be quite expensive. What is the best way to act and choose moderately inexpensive equipment that would solve our problems? Such equipment exists! For example, take two unmanaged switch models with POE ports FastEthernet and SFP port. Below in Fig. 10 are given 2 models in 4 and 8 ports, respectively: Fig .


As you can see, we are dealing with switches in the industrial form factor, which allows you to mount these models on a DIN rail and work effectively in adverse conditions.

This model belongs to the class of unmanaged switches, which in turn has a positive effect on its price. In our scheme, we can choose the UTP7204E-POE model, with four copper POE ports and one SFP port.

Let's make it clear: where is the optics connected here? And we will connect optics to the SFP module, which, in turn, will be inserted into the SFP port of the switch. Why do you need such difficulties, you ask? And we will answer - the use of various optical modules makes it possible to use this switch both in 2-fiber circuits, and in 1-fiber circuits, and on multimode fiber, and at various distances, etc., etc. In a word , select the optical module you need, insert it into the switch, and you're done!

In our case, we will choose an inexpensive optical module of the APS31123xxL2 model shown in Fig.

10 Fig. 10

This gigabit module working on 2 fibers will allow us to make connections at a distance of up to 2 km!

Well, in the cabinets we will use 4-port switches, but what can be used in the server room?

In the server room, to collect all the optical links, we need a more serious switch. So, the UTP7524GE-MX model is a gigabit modular (which is very important) managed switch. Its appearance is shown in Fig. 11

Fig. 11

It is called modular because it allows, in the process of growth of the network itself, to use additional modules for connecting optical links. In total, up to 3 units can be delivered, that is, 8 ports, 16 ports, and finally 24 ports!

Since in our case we need 6 ports to connect to 6 cabinets, then one module (see Fig. 12), for a start, will be quite enough.

Fig. 12

And, of course, we need the same optical modules that we used in the cabinets for him.

There was one question that I promised to answer: why the LC / UPC connector? Just because, as you have noticed, this connector is most often used in optical SFP modules.

Also popular now: